What Heat Input Actually Controls

Why this matters

Heat input is the one welding number that gets treated as a description of the weld. It is not. It is energy per unit length, and two procedures can hit the same figure with completely different currents, travel speeds and bead shapes, producing joints that behave differently and fail differently. A heat input limit on a procedure is doing one specific job - governing how fast the joint cools and how much the grain coarsens - and it is silent on where the energy actually went. Shops get into trouble in both directions: treating the number as a quality specification, and treating a procedure change that preserves the number as no change at all.

Before measuring anything on a live welding circuit: read current with a clamp meter around a single lead with the lead intact, and never break a welding connection, change polarity or swap a lug while the machine is energised, because opening a high-current circuit under load draws an arc at the connection. Welding leads, connectors and their insulation are covered by 29 CFR 1910.254 for arc welding equipment, and a lead with damaged insulation comes out of service rather than being taped. Work on the machine's supply side is electrical work under 29 CFR 1910.333(b)(2), which is where the general-industry lockout standard at 29 CFR 1910.147 sends electrical utilization work through its own carve-out. Any parameter trial means running a test coupon, so it carries the arc's ultraviolet at the minimum protective shade from OSHA's filter-lens table at 29 CFR 1910.133(a)(5), the fume of whatever base metal and coating are on the coupon as an inhalation route needing local exhaust at the arc or a respirator under a written program meeting 29 CFR 1910.134, and the fire prevention and fire watch the hot work permit governs under 29 CFR 1910.252(a) or 29 CFR 1926.352 in construction. Timing a pass puts a second person near the arc, so screen it.

Arc energy is not heat input, and the difference is a factor

The arithmetic that turns machine settings into energy per length - volts times amps times 60, divided by travel speed in inches per minute - is stated in the card on what a welded joint changes about the parent metal, and this card uses it rather than re-deriving it. What that formula gives you is arc energy: the energy the arc produced, not the energy the plate absorbed.

Heat input is arc energy multiplied by a thermal efficiency factor for the process. Some of the arc's energy goes into radiation, spatter and vaporised metal and never enters the work, and how much depends on whether the arc is buried under flux or open to the room. Commonly used factors put submerged arc near 1.0, the covered-electrode and wire processes near 0.8, and gas tungsten arc near 0.6. Those are process constants, not universal ones, and the values your procedure must use come from the standard the procedure is qualified to, in the edition that standard is invoked in.

This is a re-basing, not an adjustment. An arc energy figure already contains the assumption that every joule reached the plate. Applying the factor does not add a correction on top of a correct number, it replaces a number describing the arc with a number describing the joint. The consequence shows up the moment two processes are compared: a gas tungsten arc root pass and a flux-cored fill quoted at the same arc energy are not putting the same heat into the plate, because 0.8 against 0.6 is a factor of about 1.33, so the fill pass delivers roughly a third more into the work than the raw comparison implies.

What heat input does control

Every one of these holds only at constant plate thickness, joint geometry and preheat, because those set how fast the heat leaves and the heat-affected-zone card owns that half of the problem.

• Cooling rate through the transformation range, and therefore the hardness of the band beside the weld.
• The width of the heat-affected zone. More energy per inch means a wider band of parent metal taken through the damaging temperatures.
• Grain coarsening next to the fusion line, and therefore toughness in the least tough material in the assembly.
• Time spent in a temperature range that matters to the alloy, which is why the direction of the trade reverses between hardenable carbon steel and austenitic stainless. That reversal is stated and worked in the card on what a welded joint changes about the parent metal, and it is cited here rather than restated.
• The volume of metal deposited per unit length, since the pool has to be fed to be that size.

What it does not control

• Penetration depth. Penetration tracks current far more than it tracks energy per inch. Cut the current and double the travel time and the number holds while the arc stops reaching the root.
• Fusion at the sidewall. Where the arc was pointed decides that, and the lack-of-fusion card works through why a full, well-shaped bead can sit on an unmelted face.
• Bead shape. Voltage sets arc length and therefore bead width; current sets depth. Different pairs give the same product and different cross-sections.
• Freedom from defects. No heat input figure prevents porosity, undercut or a crack.

One gate, two joints that answer it oppositely

The gate: a procedure change that holds heat input constant is a change to bead shape and penetration, and it must be judged on those, not on the number that did not move.

Take two procedures on the same machine and the same wire.

• Procedure A: 26 volts, 200 amps, 12 inches per minute. Arc energy is 26 x 200 x 60 divided by 12, which is 312,000 divided by 12, or 26,000 joules per inch.
• Procedure B: 26 volts, 300 amps, 18 inches per minute. Arc energy is 26 x 300 x 60 divided by 18, which is 468,000 divided by 18, or 26,000 joules per inch.

Identical. Apply the 0.8 factor for the wire process and both land at 20,800 joules per inch of heat input, so the cooling rate, the heat-affected-zone width and the grain coarsening are, to the accuracy this arithmetic supports, the same for both. Anything a heat input limit governs is satisfied identically by either.

Now the currents: 300 against 200 is 1.5 times the current, and penetration follows current. Procedure B digs; Procedure A spreads.

Joint one: a heavy restrained groove weld, filling passes. Procedure B's deep narrow bead is the problem here. A bead whose depth exceeds its width, at constant restraint, is the classic geometry for centreline solidification cracking, because the grains grow inward from both sidewalls and meet in a plane down the middle of the bead with the last liquid trapped between them. That mechanism belongs to the cracking card. On this joint the answer is Procedure A, and the fact that the heat input number is unchanged is exactly why nobody notices the substitution.

Joint two: an open-root groove on the same thickness, root pass. Here Procedure A is the problem. The arc has to reach the bottom of the joint and melt both root faces, and 200 amps spread over a slow pass produces a large cool pool that bridges the root without fusing it. The answer is Procedure B, and again the heat input number is unchanged, so a procedure written only to an energy limit permits the wrong one.

Same number, opposite correct choices, one joint apart. That is the whole point of the gate: the energy figure is real and useful for what it governs, and it is not a description of the weld.

Correction, printed: the constants and the calculation disagree, and the calculation does not win. If a procedure states a maximum heat input and separately states an amperage range, and your calculated figure passes the energy limit while your current sits outside the stated range, the procedure has not been met. A qualified range on a parameter is a qualification boundary, and a derived number that happens to land inside a limit does not requalify a parameter that fell outside its own. The procedure governs; the arithmetic is a check on it.

Check against the sibling rules, with the figures printed. The preheat card states that cooling rate depends on chemistry, combined thickness and hydrogen level together; this example holds thickness and geometry constant between A and B by construction and makes no preheat claim, so it does not use heat input to substitute for a preheat determination. The lack-of-fusion card names low-energy conditions at the root as a fusion risk; that is the reason Procedure A is rejected on joint two, at 200 amps against B's 300, rather than on any energy figure, since both are 26,000 joules per inch. The cracking card owns solidification cracking; the depth-to-width geometry is cited to it and the 1.5 times current ratio is what is claimed here.

What would change the answer. Move to submerged arc and the efficiency factor moves toward 1.0, so the same arc energy delivers about 25 percent more into the plate than the 0.8 processes did, and a heat input ceiling written for a wire process is exceeded without the machine settings changing at all. Move to a thin section and the whole comparison shifts, because thin material cools through fewer paths and the same energy per inch produces a much slower cooling rate. Move to austenitic stainless and the direction of the trade reverses, which is the parent-metal card's finding, not a new one here.

How to verify the number you are quoting

• Measure travel speed, do not estimate it. Mark a known length on the coupon, time the pass with a stopwatch, and convert: 6 inches in 30 seconds is 60 times 6 divided by 30, which is 12 inches per minute. Estimated travel speed is the largest error in most shop heat input figures and it sits in the denominator, where a 20 percent estimate error is a 20 percent error in the result.
• Read amperage and voltage while welding, from a meter, not from the dial. A dial setting is a demand and the machine's output under load is the quantity in the formula. Clamp the lead intact rather than opening the circuit.
• Say which figure you are quoting, every time. Arc energy and heat input differ by the efficiency factor, and a number handed over without its label is a number the next person will apply with a different assumption.
• State the process with the figure. Without the process there is no efficiency factor, and 26,000 joules per inch means three different things depending on which arc produced it.
• Compare only your own procedures on the same thickness and joint. Two shops quoting heat input on different geometries are not comparing anything, because the plate, not the arc, sets what the number does.

References

• AWS D1.1 structural welding code for steel, in the edition invoked by your project specification or the adopted building code, for heat input provisions, qualified parameter ranges and the process efficiency factors that apply to them
• ASME Boiler and Pressure Vessel Code Section IX with the applicable piping or vessel code, as adopted by your jurisdiction, for procedure qualification variables where the work is pressure-retaining
• 29 CFR 1910.254 for arc welding and cutting equipment, leads and connections; 29 CFR 1910.333(b)(2) for work on the electrical supply side
• 29 CFR 1910.133(a)(5) for filter shade; 29 CFR 1910.134 for respiratory protection; 29 CFR 1910.252(a) and 29 CFR 1926.352 for hot work fire prevention on a test coupon
• See related: What a Welded Joint Changes About the Parent Metal; What the Heat-Affected Zone Is and Why It Governs; What Lack of Fusion Is and Why It Hides